背景：黑色素瘤是一種高度侵襲性的皮膚癌，一旦發生轉移，預後極差。鑑別具有轉移抑制潛力的生物標誌物，對於改善臨床治療成效至關重要。材料與方法：本研究整合三個GEO資料庫數據集（GSE65904、GSE59455、GSE8401），分析轉移性與原發性黑色素瘤組織間的差異表現基因（DEGs）。藉由STRING、Cytoscape（MCODE、CytoHubba）及Venn分析進行GO/KEGG功能富集、蛋白質交互作用（PPI）網絡建構與核心基因篩選。臨床相關性則透過GEPIA與OSdream平台加以驗證。結果：共鑑定出37個重疊的DEGs，富集於角質形成細胞分化與上皮屏障功能等相關途徑。其中，loricrin（LOR）、角蛋白6B（KRT6B）與filaggrin（FLG）在黑色素瘤中表現一致下調，並被鑑定為核心基因。LOR 表現量與較長的無轉移存活期（metastasis-free survival, p = 0.0466）相關，且其表現與ADAM12、ITGA4、CDK1等促轉移基因呈負相關。此結果顯示LOR可能具轉移抑制作用。結論：LOR 為一潛在預測黑色素瘤轉移與疾病進展的生物標誌物，具臨床預後與治療開發應用價值。

Background: Melanoma is a highly aggressive skin cancer with poor prognosis once metastasis occurs. Identifying metastasis-suppressive biomarkers is crucial for improving clinical outcomes. Materials and methods: We integrated three GEO da-tasets (GSE65904, GSE59455, GSE8401) to identify differentially expressed genes (DEGs) between metastatic and primary melanoma tissues. GO/KEGG enrichment, PPI network construction, and hub gene analysis were performed using STRING, Cytoscape (MCODE, CytoHubba), and Venn analysis. Clinical significance was val-idated via GEPIA and OSdream platforms. Results: We identified 37 overlapping DEGs enriched in keratinocyte differentiation and epithelial barrier functions. loric-rin (LOR), Keratin 6B (KRT6B), and filaggrin (FLG) were consistently downregu-lated in melanoma and identified as hub genes. Among them, LOR expression corre-lated with longer metastasis-free survival (p = 0.0466) and was negatively associated with ADAM12, ITGA4, and CDK1 expression in melanoma. These results suggest that LOR may suppress melanoma metastasis. Conclusion: LOR is a potential bi-omarker for predicting melanoma metastasis and progression, highlighting its value for prognosis and therapeutic targeting.